Weapons of Mass Production: <$1,000 Camera

Posted by Kevin Good on 01/22/13

The cameras were:

Canon T4i with 18-55mm kit lens ($899)
Sony NEX-6 with 16-50mm kit lens ($998)
Panasonic GH2 with 14-42mm kit lens, both stock and with Flowmotion v2.02 hack ($699)

Let's get this out of the way: All three cameras are very capable, great cinematic images can be shot with any of them. But they're not quite up in that stratosphere of "so good it just doesn't matter any more". They each have their strengths & weaknesses.

I know there are a lot more cameras I could've chosen from. I chose these because (i) Canon has consitently been a leader in DSLR video quality, so I felt they should be represented, (ii) mirrorless is the way of the future for video, so I wanted the others to be mirrorless, (iii) these all have larger chips that 'traditional' video cameras (the handycam you'd get at Best Buy made specifically for video) which gives us better light sensitivity, shallower depth-of-field, and that cinematic look that viewers of my show are concerned with.

Canon T4i:

I chose this one as the overall winner even though it has a couple shortcomings. It's not as sharp as the Panasonic, and the form factor (with optical viewfinder instead of electronic) isn't as good for video as the others. But it was the best in low light, it was the best with dynamic range, it was just about even on the other tests, and there is a hack available for it (Magic Lantern) that gives some really nice functionality. It worked predictably, it has a nice articulating touch-screen, it has a histogram to aid exposure, and having the cleanest image matters a lot to me. I think it's embarrassing that it can't resolve the same detail as the much older Panasonic, but you have to remember that these cameras are all stills cameras first, so sometimes the video functionality comes as an afterthought.

Sony NEX-6:

The Sony actually did quite well, and if size (for travel) is a concern, I'd suggest this one. It's just a hair behind I think in the image metrics, but not in a huge way. The focus issue did drive me crazy. Maybe there's a menu option I didn't see that prevents it from going to sleep. I actually had to reshoot the sharpness (helicopter) comparison because it did that same focus thing to me. That's why when it did it again for the dancing, I decided to leave it in the video that way. It has focus peaking (that shows you a highlighted version of in-focus areas on the display) which really helps with video, and I doubt this issue would ever come up when shooting stills. It's a solid camera, very similar to Canon in image quality, great if you're a space-conscious traveller.

Panasonic GH2 with Flowmotion Hack:

The Panasonic is pretty plain and simple: It is the sharpest. In fact I think it's the sharpest camera under $10,000. I think this camera is the clearest proof that the manufacturers are holding-out on us, and releasing cameras that aren't as aggressively great as they could be, because they want to segment their market and control the upgrade timeline. I was actually surprised as how sharp the camera is right out of the box, sans hack. The hack really comes into play when you have a lot of motion in the scene. On the stock version, the compression falls apart into jpeg-y blocks, on the hacked version it's clean throughout. Sadly, the camera lost (sometimes emphatically) on every other test. So even though it's the cheapest and sharpest by some margin, I can't give it my whole-hearted recommendation.

0 comments to "Weapons of Mass Production: <$1,000 Camera"

<<Previous Post

Weapons of Mass Production: Action Cam Showdown

The F%^&ing Nikon D800 vs. Canon 5D mkIII Shootout Part II

Posted by Kevin Good on 05/18/12

This is part 2 of my previous test. It is even more idiotic, but I find the results very interesting, particularly for stills shooters. Pedantic analysis below.

Test 7: Got Bits?

This is my favorite test. I might have to come up with a normalized (repeatable) version of this. Because I think this test, more than the others, shows the quality of the image for processing. I do a lot of post, and frequently I say to myself "I just wish this was cleaner, or those colors still looked right [despite being pushed around]." The underwater picture earlier on for example was mine, and I had to run such strong noise reduction on it that there's barely anything left.

I think both of the cameras performed very well, and very similar to each other, until we get into the darks. Then obviously the Nikon holds on better. I also checked them on the higher highlights (out the window) to see how they did, but it was essentially the same as my lighter sign, so I ommitted it from the video.

In case it's not clear: Both signs had the camera brand printed full black on the paper, and the text underneath printed to be barely visible at all to the naked eye. Furthermore the words "BEEZ" and "DEEZ" were printed with each letter in a different solid color, to see if color really threw things off. It's kind-of a poor man's contrast test chart.

Test 8: Stills ISO

This one was pretty straightforward. We jacked the ISO, we shot some comparisons. I think this is where the real upset comes. There are still people out there smack-talking Nikon's 36 megapixel chip because we've all come to expect high megapixel counts to mean a noisey image. It's simply not the case, and it definitely gets the edge over the 5D3 in this case. And the Canon is no slouch! We're talking about a couple of the most sensitive cameras on the planet here. Great stuff from both.

Winners & Losers:

"But Kevin, you didn't test the video codecs, but Kevin, you didn't mention the crop mode focus points color this or that whatever! But Kevin the weights you gave these things are arbitrary! But Kevin, you talk about some things like 'weathersealing' that you didn't actually test and so it's just hearsay!"

I know. The idea was to convey an overall sense of how the cameras stack up, as per the stuff that matters to me and most shooters (but not all), and to not bore everyone to tears by getting into the minutae. I already start to snooze by the time we get to the end. And yes, the weights are arbitrarily chosen by me. But it's my video and I'm trying to suggest how important I think they are. Clean ISOs for example, is very important, but they were nearly identical in stills so it gets a light weight. In video they're hugely different so it's a heavier weight. It's also a way of visualizing what camera might be right for YOU by imagining the meter without a couple of the words or with a couple weighted differently. e.g. If price is no object, the Nikon gets a modest win for stills and gets majorly clobbered for video. We're all different shooters. And while this segment is seriously flawed scientifically-speaking, I felt it would still be useful for a lot of people.

If you have thoughts about other important things I missed or misrepresented, comment-away; this isn't conclusive. But I think it's a good way of getting people up-to-speed quickly on the discussion.

Visible Spectrum:

Ah-mazing.

I did go on later to do some rudimentary tests involving the visible spectrum, and the cameras are great (as expected from most modern cameras) so there isn't really much of a need for a serious version of this test any more.